Patents and the Funding Gap

A patent can do several things for a startup. It creates a time-limited right to exclude others (not an affirmative right to make or sell, and not proof of freedom to operate), it can signal technical credibility to investors, and in some circumstances it can help a company raise equity or debt.

Its value is not universal: it varies by technology, claim scope, stage, business model, and the alternatives a company has. This page brings together three separate bodies of research: the effect of first-patent approval on startup financing, gender differences during patent prosecution, and gender differences in serial venture entrepreneurship. They are related, but they are not one identified causal chain. We separate what each study directly establishes from what can reasonably be inferred across them.

This article presents published research data for educational purposes. It is not legal, policy, financial, or investment advice. Patent gender statistics depend on name-based inference and methodology choices: see How we measure.

What a first patent can do for a startup

The cleanest causal evidence comes from Farre-Mensa, Hegde and Ljungqvist, "What Is a Patent Worth? Evidence from the U.S. Patent Lottery" (Journal of Finance, 2020). Studying 34,215 first-time U.S. startup applicants, the authors exploit quasi-random differences in examiner leniency to estimate the effect of obtaining approval separately from the quality of the underlying invention.

Approval raised a startup's chances of securing venture-capital funding over the next three years by about 47% (a relative increase), and startups whose first application was approved went on to show, over five years, roughly 55% higher employment growth and 80% higher sales growth than comparable startups whose application was not approved. These are cumulative growth differences, not simply "55% more employees." Improved access to external financing (VC, debt, and public capital) is one important channel the authors identify; it is not established as the sole mechanism behind the growth effects.

A patent is also a signal

Farre-Mensa measures what a patent does. A second line of research explains part of why. For a young company with little revenue, a patent is a credible quality signal to investors. Hsu and Ziedonis (Strategic Management Journal, 2013), studying 370 venture-backed semiconductor startups, found that firms with a larger patent stock were more likely to attract initial funding from a prominent VC, and that the effect was strongest for founders without a prior track record. A patent can stand in for a reputation a first-time founder has not yet built.

The software caveat, stated plainly. The financing effect is not uniform across every patent or sector, and as a software-adjacent company we should be the first to say so. Two studies sharpen the picture for software. James Hicks (Northwestern University Law Review, 2024) examined business-method software startups and found no evidence that patents channeled investment to them, nor better downstream outcomes like acquisition or IPO. Helmers and Love ("Patent Protection and Software Firm Financing," 2025), using a difference-in-differences design around the 2014 Alice decision that restricted software patentability, found software startups more exposed to the restriction fared no worse on financing, growth, or exit than less-exposed ones.

Neither overturns Farre-Mensa (whose historical sample actually found a larger effect for IT firms than biotech). They use different designs and treatments; together they show the financing value of software patents is not universal and not fully settled. The honest version is not "any patent gets funded." Credible patent rights can function as a financing signal, the size of that effect depends on sector, stage, and claim quality, and patent strategy should be evaluated company by company rather than treated as an automatic fundraising formula.

Gender differences across patent prosecution

Women face documented disparities in patent outcomes, though the exact stage-by-stage magnitudes and causes are still debated. Gu, Lewellen, Mao and Qin ("To Grant or Not to Grant: Inventor Gender and Patent Examination Outcomes," working paper, November 2025; 988,125 U.S. applications, 2001-2017) report that all-women teams are about 20.9 percentage points less likely to have a patent eventually granted than all-men teams (against a full-sample grant rate of 68.5%). About 5.4 points remain after controls for narrow technology subclass, examiner, entity, and inventor characteristics. The raw gap partly reflects that women are more concentrated in technology areas with lower approval rates, so the 20.9 points is not itself evidence that examiners caused the whole difference, and 5.4 is an adjusted association, not a causal decomposition.

The rare-name analysis is a comparison, not an experiment. In an observational analysis of solo-inventor applications, the authors compare inventors with common versus rare first names (assuming rare names are harder for an examiner to code as male or female). The first-action difference becomes statistically indistinguishable from zero among the rare-name group. That is suggestive of a gender-signaling mechanism, but no application was resubmitted under a different name and nothing was redacted or randomized. Treat it as their estimate, not a settled fact. (Fuller treatment on Where women's patents disappear.)

The examiner-bias literature is genuinely unsettled. A separate 2025 working paper (Avivi, "Are Patent Examiners Gender Neutral?") uses patent-text embeddings to control for application content and finds no average content-adjusted first-action difference between mixed-gender and all-men teams, while finding meaningful variation across individual examiners. Research consistently finds gender differences in ultimate outcomes, but the existence, average size, and causes of a first-action examiner effect remain contested.

The strongest newer evidence is about help, not just bias. A 2025 randomized trial at the USPTO (Pairolero et al., AEJ: Economic Policy) found that extra assistance for unrepresented applicants improved patent-obtainment for both men and women, with the treatment effect more than 12 percentage points larger for women, largely through successful examiner amendments. And Aneja, Reshef and Subramani (Review of Economics and Statistics, ~1M applications) found applications listing women 3.6 to 6.9 pp less likely to be continued after an early rejection, estimating that differential abandonment accounts for more than half of the issued-patent gap in their sample. Lower continuation is a resource-and-support problem (an employer, assignee, or counsel often controls the budget), not an individual "confidence" defect.

The funding gap compounds after entry

Venture finance has its own compounding mechanism, separate from patents. Hebert, Yimfor and Tookes, "Financing the Next VC-Backed Startup: The Role of Gender" (NBER working paper, 2025), report that women are 13.3% of VC-backed founders but only 4% of founders who start three or more VC-backed companies.

Comparing men and women who co-founded the same prior company, the authors find that following a failure, women are 22.5% less likely than their male cofounders to found another VC-backed startup. Among those who return, women raise 53.3% less capital after a prior failure and 24.6% less after a prior success. They interpret this as consistent with unequal treatment in the capital market.

This shows funding disadvantages can accumulate across a founder's career, even after she has entered the VC-backed ecosystem. Importantly, the study does not test patents as the cause of these disparities. We present it as a parallel compounding mechanism, not a downstream effect of the patent gap.

Established links versus inferred links

The honest way to read this page is as an evidence map, not a single causal chain. Some links are directly estimated; the connection between them is plausible but not measured.

Where ObviouslyNot fits, and where it does not

One actionable upstream place to intervene is the earliest one: making sure inventive work becomes a reviewed disclosure in the first place. ObviouslyNot's scanner reads a codebase and surfaces candidate technical concepts, each with traceable references back to the code, for human review. It offers an additional route into an internal invention-harvesting process, without requiring an engineer to arrive with a fully formed patent idea or self-identify as an inventor.

The limits matter, and we state them:

• It does not determine inventorship. Under U.S. law, inventorship turns on a human contribution to the conception of claimed subject matter. Writing code, making commits, or completing a disclosure form does not by itself make someone a legal inventor; someone who made a conceptual contribution may be an inventor even if they wrote little code. Only natural persons can be inventors. Counsel decides inventorship claim by claim.
• "Strategic concept" is a workflow term, not a legal category. The scanner does not assess patentability, novelty, ownership, validity, freedom to operate, commercial value, or the likelihood of financing.
• Identity-blind is not automatically neutral. Not filtering by who wrote the code removes one bias, but results still reflect who was assigned to which repositories and whose non-code contributions are visible. It is a design choice to reduce reliance on self-nomination, not proof of a fair outcome.
• Keep it confidential. Handle scan outputs and source code under confidentiality controls and consult counsel before any public release; prefiling public disclosure can forfeit rights in many jurisdictions.

The honest claim is a design hypothesis, not a demonstrated outcome: systematic, code-based concept discovery may reduce reliance on self-nomination at one early stage of invention review. Whether it broadens participation, produces more valuable disclosures, or narrows any gender gap is an empirical question that needs outcome data.

Sources

More from the patent gender gap series